The observations of ripples in the plasmapause surface, of attached plasmatails and detached cold plasma elements are recalled. The mechanisms proposed for the formation of plasmatails and of detached plasma elements are also recalled. In this contribution we describe an alternative mechanism which fits currently available observations. This new physical mechanism is based on the enhancement of the day-night asymmetry of the plasmapause during substorms, and the formation of a bulge in the dayside plasmasphere, as first observed by Gringauz and Bezrukikh (1976). Eastward magnetospheric convection shears this bulge because its tip is convected with a smaller azimuthal velocity than the plasma closer to Earth. As a result of this shear motion an attached plasmatail develops in the afternoon-dusk sector. The westward edge of this plasmatail corresponds to the well defined and sharp 'shoulder' in the equatorial plasmapause surface observed by Carpenter (1970) in whistler data, and by Higel and Wu (1984) with GEOS-2 measurements.